sign in sign up
<<
Home , Transactivation, Transcription factor

Proc. Natl. Acad. Sci. USA Vol. 94, pp. 8485–8490, August 1997 Biochemistry

Interaction of the human androgen function with the general factor TFIIF

IAIN J. MCEWAN* AND JAN-ÅKE GUSTAFSSON

Department of Biosciences, Novum, Karolinska Institute, S-141 57 Huddinge, Sweden

Communicated by Elwood V. Jensen, University of Hamburg, Hamburg, Germany, May 27, 1997 (received for review January 28, 1997)

ABSTRACT The human (AR) is a tion factors, and thus the polymerase, to the (re- -activated that regulates im- viewed in refs. 17–19). This can be achieved by direct contact portant for male sexual differentiation and development. To between the and the general transcription factors better understand the role of the receptor as a transcription and͞or interactions by means of (refs. 17 factor we have studied the mechanism of action of the N- and 19–21 and references therein). terminal transactivation function. In a –protein inter- In recent years a number of interactions have been described action assay the AR N terminus (amino acids 142–485) between members of the – receptor selectively bound to the basal transcription factors TFIIF and superfamily and basal transcription factors and co–activator the TATA-box-binding protein (TBP). Reconstitution of the proteins (see ref. 22 and references therein). However, very transactivation activity in vitro revealed that AR142–485 fused to little is known concerning the identity of interacting proteins the LexA protein DNA-binding domain was competent to with the human AR. To better understand the mechanism of activate a reporter in the presence of a competing DNA gene regulation by the human AR we have screened a panel of template lacking LexA binding sites. Furthermore, consistent general transcription factors for binding to the receptor N- with direct interaction with basal transcription factors, ad- terminal and have reconstituted recep- dition of recombinant TFIIF relieved squelching of basal tor-dependent activation under -free conditions. A region transcription by AR142–485. Taken together these results sug- of the N terminus, containing the major transactivation activ- gest that one mechanism of transcriptional activation by the ity, is capable of recruiting the general transcription machinery AR involves binding to TFIIF and recruitment of the tran- to a target promoter and shows selective binding to the general scriptional machinery. transcription factor TFIIF.

The androgen receptor (AR) is a member of the steroid– MATERIALS AND METHODS thyroid superfamily and mediates the ef- fects of the male sex testosterone and dihydrotest- AR Expression Constructs. The DNA sequence coding for osterone (for review see ref. 1). in the receptor amino acids 142–485 of the human AR N terminus was proteins have been identified in disorders of male sexual amplified by using the Expand Long Template PCR system differentiation (2, 3), X-chromosome-linked spinal bulbar (Boehringer Mannheim) from plasmid pSVARo (a gift from muscular atrophy (4, 5), prostatic carcinoma (6, 7), and male A. O. Brinkmann, Erasmus University, Rotterdam, The Neth- (8). Although there is good evidence that the AR erlands; see ref. 9). The primers used were ARN142, 5Ј- binds to DNA response elements and activates gene expres- GCGCGCAGATCTCTGCCGCAGCAGCTGCCAGC-3Ј, sion, the underlying mechanisms are not well understood. The and ARC485, 5Ј-GCGCGCGGATCCGCTTTCCTGGC- C-terminal steroid-binding domain and the central DNA- CCGCCAGCCCC-3Ј. The PCR product was cleaved with binding domain show significant homology between ARs of BamHI and BglII and ligated into pET-19bm (23) previously different species and also with other members of the nuclear digested with BamHI. The resulting plasmid, pET-AR4, ex- receptor superfamily (ref. 1 and references therein). In con- pressed the AR142–485 domain fused to an N-terminal histidine trast, the N terminus of the protein is more divergent and is tag. The insert was checked for orientation and sequenced characterized by homopolymer tracts of glutamine, glycine, (Dye terminator , Perkin–Elmer). The ex- and proline residues (ref. 9 and references therein). Regions pression plasmid pET-AR4-Lex was constructed by ligating a within the N terminus of the human and rat receptors impor- fragment encoding the LexA DNA-binding domain (LexADBD, tant for transactivation have been delineated by deletion amino acids 1–87) into the regenerated BamHI site, 3Ј of the analysis (10–13), the use of fusion proteins (14), and point AR sequence in pET-AR4. mutations (15). These studies have highlighted the region Expression and Purification of Recombinant Proteins. between amino acids 142 and 370 (numbering for the human AR142–485 with or without LexADBD was expressed in Esche- receptor), although sequences both N-terminal and C-terminal richia coli strain BL21 plys by isopropyl ␤-D-thiogalactoside of this region appear to play an important role in the full (IPTG; 1 mM) induction, and the recombinant proteins were activity of the wild-type AR and͞or in promoter specific purified from the soluble fraction by Ni2ϩ–nitrilotriacetate activity (see ref. 14). (NTA) affinity chromatography. The bound protein was eluted Transcription of mRNA coding genes involves the concerted with 200 mM imidazole and dialyzed against 25 mM Hepes, pH action of RNA polymerase II and a set of at least five general 7.6͞100 mM sodium acetate͞1mMDTT͞0.01% Nonidet P-40. transcription factors (see refs. 16–19 for recent reviews). One Recombinant yeast TATA-box-binding protein (TBP) and mechanism by which gene regulatory proteins are thought to human TFIIF (RAP30 and RAP74) were expressed in bacteria function is by recruiting one or more of the general transcrip-

Abbreviations: AR, androgen receptor; AR142–485, AR residues 142– The publication costs of this article were defrayed in part by page charge 485; LexADBD, LexA DNA-binding domain (residues 1–87); TBP, TATA-box-binding protein; SRF, . payment. This article must therefore be hereby marked ‘‘advertisement’’ in *To whom reprint requests should be addressed at present address: accordance with 18 U.S.C. §1734 solely to indicate this fact. Department of Molecular and Cell , University of Aberdeen, © 1997 by The National Academy of Sciences 0027-8424͞97͞948485-6$2.00͞0 Institute of Medical Sciences, Foresterhill, Aberdeen AB25 2ZD, PNAS is available online at http:͞͞www.pnas.org. Scotland, United Kingdom. e-mail: [email protected].

8485 Downloaded by guest on October 2, 2021 8486 Biochemistry: McEwan and Gustafsson Proc. Natl. Acad. Sci. USA 94 (1997)

and partially purified as described previously (24). Protein concentrations were measured against BSA standards using the Bradford reagent (Bio-Rad). Protein–Protein Interaction Assay. The microtiter plate interaction assay was essentially as described previously (23, 24). Briefly, AR142–485 or BSA control in binding buffer [20 mM Hepes, pH 7.6͞10% (vol͞vol) glycerol͞100 mM KCl͞0.2 mM EDTA͞5 mM MgCl2͞5 mM 2-mercaptoethanol] were allowed to adsorb to the surface of a scintillation-microtiter plate (Wallac, Oy, Finland). Unoccupied surfaces were sub- sequently blocked with binding buffer containing 5 mg͞ml BSA, and the wells were incubated with binding buffer con- taining 1 mg͞ml BSA and radiolabeled human basal transcrip- tion factors, synthesized in a rabbit reticulocyte lysate system (Promega). After extensive washing with binding buffer ϩ 1 mg͞ml BSA, the bound radioactivity was measured directly in a micro ␤ counter (Wallac, Oy, Finland), and bound proteins were recovered in SDS sample buffer. In Vitro Transcription and Squelching Assays. Preparation of yeast nuclear extracts for in vitro transcription, together with reporter genes and reaction conditions have all been described in detail previously (24–26).

RESULTS The N-Terminal Region of the Human AR Contacts Basal Transcription Factors. The N terminus of the AR has been shown to contain a complex transactivation function made up of multiple regions (see Introduction). In an attempt to understand the mechanism by which the AR activates tran- scription, a panel of basal transcription factors was screened for interactions with the receptor transactivation domain. A

FIG.2. (A) Screening of 35S-labeled human general transcription factors for binding to AR142–485. The measured radioactivity is plotted relative to wells containing BSA only, which was set at 1. The binding of TBP to the C-terminal transactivation domain of the herpes simplex viral activator protein VP16 is shown for comparison. The results represent the mean Ϯ SD of at least four observations from two or more independent experiments, except TFIIH, subunits where the mean of two observations only is shown. (B) SDS͞PAGE analysis of the bound RAP74 subunit of TFIIF to AR142–485. The input represents 5% of the material incubated per well.

polypeptide containing amino acids 142–485 of the human receptor was expressed and purified by metal chelation chro- matography (Fig. 1). The purified protein was allowed to adsorb onto the surface of a microtiter plate and incubated with 35S-labeled basal transcription factors TFIIB, TBP, TFIIE␣ and -␤, TFIIF (RAP30 and RAP74), and two of the subunits of TFIIH (p44 and p62). AR142–485 interacted selec- tively with the RAP74 subunit of TFIIF and showed modest binding to the RAP30 subunit of TFIIF and to TBP (Fig. 2A). Little or no significant binding was observed with TFIIB, TFIIE, or the two subunits of TFIIH, as judged relative to the BSA control and binding to proteins lacking a transactivation function (Fig. 2A and data not shown; see also refs. 23 and 24). To evaluate the significance of these observations, we com- pared the binding of AR142–485 with previously reported in- teractions between the serum response factor (SRF) and RAP74 (27) and the potent viral activator VP16 and TBP (see ref. 16). The binding of AR142–485 to RAP74 was as strong, if not stronger, compared with SRF245–508 (see references 23 and 24; also data not shown). While the interaction with TBP was FIG.1. (A) Schematic representation of the human AR showing the domain organization and the region of the N terminus used in the comparable to that seen with VP16 (Fig. 2A and refs. 23 and present study. (B) Purification of His-tagged AR142–485 and 24). Recovery of the bound material and analysis by SDS͞ AR142–485-Lex proteins. Proteins were estimated to be at least 75% PAGE confirmed that the measured radioactivity represented pure from the Coomassie blue-stained gel. intact labeled RAP74 binding to AR142–485, while little or no Downloaded by guest on October 2, 2021 Biochemistry: McEwan and Gustafsson Proc. Natl. Acad. Sci. USA 94 (1997) 8487

labeled protein was recovered from the BSA-only control (Fig. DNA 2). This activation was dependent on the presence of the 2B). Thus, a region of the AR N terminus containing the AR142–485 polypeptide, as LexADBD alone showed no stimula- transactivation function bound selectively to the basal tran- tion of transcription (Fig. 3 B and C; DNA 2). Furthermore, scription factors TBP and TFIIF. transcription from the AdML⌬-50 template was essentially Reconstitution of Receptor Transactivation Activity Under eliminated in the presence of AR142–485-LexADBD (Fig. 3B; Cell-Free Conditions. To study the mechanism of transcrip- DNA 1). Since the 1xLexCG reporter template (DNA 2) was tional activation and to evaluate the functional relevance of the added after preincubation of the competing template (DNA observed interactions, the N-terminal transactivation activity 1), these results suggest that the AR functions, at least in part, was reconstituted under cell-free conditions. The AR142–485 to recruit the general transcription machinery to a target region was fused to the heterologous DNA-binding domain of promoter. the bacterial LexA protein (LexADBD) and was expressed and Previously, we have used a squelching assay to assess the role purified (Fig. 1). Previously, a relatively crude preparation of of protein–protein interactions in activator function (23–25), full-length human AR, expressed using vaccinia virus, was where the squelching is thought to result from the competition shown to activate a in HeLa nuclear extracts for a limiting pool of target factor(s) (31). Addition of (28). However, the mechanism(s) of activation was not ad- AR142–485, in the absence of the DNA-binding domain, to- dressed in this study. To test whether AR142–485 could recruit gether with a basal reporter gene resulted in a concentration- factors to the promoter, the basal transcription machinery was dependent inhibition of transcription, consistent with the allowed to assemble on a competing template, prior to addition transactivation domain interacting with a component(s) re- of the reporter DNA and the receptor. In the absence of quired for basal transcription (Fig. 4). This is likely to be a recruitment, transcription from the first template would be specific effect, since the squelching occurred within the con- expected to occur preferentially. Therefore, nuclear extracts centration range required for activation and was comparable were preincubated with the AdML⌬-50 basal promoter re- to that seen with other activator proteins, VP16 (32, 33), porter gene (DNA 1), prior to the addition of the 1xLexCG receptor (GR)-␶1 (25), c- (24), and the reporter gene (DNA 2), NTPs, and activator (Fig. 3A). Under arylhydrocarbon receptor (AhR) and the AhR nuclear trans- these conditions the levels of basal transcription from DNA 2 locator protein (ARNT) (23) transactivation domains under were 10–20% lower than observed in the absence of DNA 1 similar conditions. (Fig. 3B lane 1 and data not shown). Addition of AR142–485- Recombinant TFIIF Reverses AR-Dependent Squelching. LexADBD resulted in an approximately 5-fold activation of The squelching assay provides a functional test for the signif- transcription relative to no added activator (Fig. 3 B and C; icance of interactions identified in protein–protein binding

FIG.3. (A) Overview of the experimental protocol (time shown in minutes) followed to reconstitute the transactivation activity of the AR142–485 domain fused to the LexADBD in yeast nuclear extracts. DNA 1 represents the adenovirus major late basal promoter upstream of the 180-bp G-free cassette (AdML⌬-50; ref. 29), and DNA 2 is the CYC1 basal promoter with one LexA binding site cloned upstream fused to the 380-bp G-free cassette (1xLexCG; ref. 23). (B) Representative experiment showing RNase T1-resistant transcripts derived from 1 and 2 in the absence or presence of 50 pmol of LexADBD or AR142–485-LexDBD. Note two specific transcripts are obtained from DNA 2 (see ref. 30). (C) Transcripts from DNA 2 were quantitated using a Bio-imaging analyzer BAS2000 (FujiFilm), and mean and SD (n ϭ 3) are plotted relative to basal transcription (no added activator) set to 1. Downloaded by guest on October 2, 2021 8488 Biochemistry: McEwan and Gustafsson Proc. Natl. Acad. Sci. USA 94 (1997)

FIG.4.(A) Representative experiment showing squelching of basal transcription from CYC1 basal promoter (p⌬CG; ref. 30) by increasing amounts of AR142–485.(B) Quantitation of basal transcrip- tion in the presence of increasing amounts of AR142–485. The results were quantitated and plotted as described in legend to Fig. 3C for three independent experiments.

studies, since the addition of putative target factors would be predicted to relieve the squelching. Fig. 5A and C shows that addition of partially purified human TFIIF resulted in an inhibition of basal transcription in the absence of AR142–485 but significantly relieved the squelching of transcription by FIG.5. (A)(Left) Relief of AR142–485-dependent squelching by the AR142–485. Relief of squelching was measured as the increase addition of recombinant TFIIF. Reversal of squelching was calculated in the ratio of relative transcription in the presence versus the as the ratio of the relative transcription in the presence (60 pmol) absence of the activator, AR142–485, relative to the amount of versus the absence (0 pmol) of AR142–485 and plotted against the TFIIF added (Fig. 5A Left). Recombinant TFIIF also relieved amount of TFIIF added (ng of each subunit). Data for at least three experiments are shown. In the absence of TFIIF subunits, the mean of auto squelching by the SRF245–508 transactivation domain but not squelching by the GR-␶1 (Fig. 5A Right). Under similar this ratio is 0.26 Ϯ 0.08 (n ϭ 6), and an increase in this ratio is indicative of reversal of squelching. (Right) Reversal of self-squelching by conditions, recombinant TBP was unable to reverse AR142–485- Gal-SRF245–508 (15 pmol) but not squelching of basal transcription by dependent squelching, although there was an overall increase in GR-␶1 (20 pmol) by 300 ng of TFIIF. (B) Addition of recombinant the levels of transcription in the absence of AR142–485 (Fig. 5B TBP fails to relieve AR142–485-dependent squelching (Left). Condi- and data not shown). In contrast, TBP successfully overcame tions as in A. In the absence of TBP the mean of the ratio is 0.25 Ϯ squelching of basal transcription by VP16 (Fig. 5B Right) and 0.11 (n ϭ 3). In contrast, TBP did reverse squelching of basal the c-myc (24) transactivation domains. Taken together, these transcription by VP16 (2–4 pmol) (Right). Mean and SD for at least data strongly suggest that TFIIF is a target factor for the AR four experiments are shown. (C) Representative experiment showing the effect of recombinant TFIIF on transcription levels in the absence N-terminal transactivation function. or presence of AR142–485. The amounts of AR142–485 (pmol) and TFIIF subunits (ng) added are shown above each lane.

DISCUSSION to the interaction with TFIIF. This lack of a functional correlation with TBP binding is not thought to be due to the Regulation of transcription by gene regulatory proteins in- use of recombinant yeast TBP in the functional assay, as yeast volves both direct and indirect interactions with the general TBP was capable of reversing squelching by VP16 (Fig. 5B) and transcription machinery (see Introduction). In the present the c-myc (24) transactivation domains under similar condi- study we show that the N-terminal transactivation function of tions. Although no significant binding of the AR142–485 domain the human AR acts, at least in part, to recruit the transcrip- to other components of the basal transcription machinery was tional machinery to the promoter. Consistent with this, the AR observed we cannot exclude the possibility that such interac- selectively binds to the general transcription factor TFIIF, and tions were masked or occur with a component not tested. this interaction is sufficient to relieve squelching of basal Wilson and co-workers (1) have previously reported that the transcription under cell-free conditions. The inability of TBP N terminus of the human AR (amino acids 1–503) squelched to relieve squelching by AR142–485 suggests that, although this activation by the full-length receptor in transient transfection region of the receptor can bind to TBP in an interaction assay, studies. However, the nature of the factor(s) involved remains in the context of a functional assay this interaction is secondary to be determined in this system. Downloaded by guest on October 2, 2021 Biochemistry: McEwan and Gustafsson Proc. Natl. Acad. Sci. USA 94 (1997) 8489

TFIIF is composed of two subunits, RAP30 and RAP74, Brinkmann (Erasmus University), Z. F. Burton (Michigan State that were originally identified as RNA polymerase II- University), J.-M. Egly (Institut de Ge´ne´tique et de Biologie Mole´cu- associated proteins (RAPs; reviewed in ref. 34). Furthermore, laire et Cellulaire, Institut National de la Sante´et de la Recherche TFIIF has been identified as a component of the recently Me´dicale),S. Hahn (Fred Hutchinson Cancer Research Center), M. described holo-RNA polymerase complex found in Hampsey (Louisiana State University Medical Center), R. Kornberg ͞ (Stanford University), R. Prywes (Columbia University), D. Reinberg both yeast (35, 36) and mammalian (37, 38) cells. TFIIF is (University of Medicine and Dentistry of New Jersey–Robert Wood unique in that it functions at two discrete steps in the tran- Johnson Medical School), and R. G. Roeder (Rockefeller University). scription process: first, during transcription initiation, TFIIF is We thank Drs. Anthony Wright and George Kuiper (Department of responsible for selective recruitment of the RNA polymerase Biosciences, Karolinska Institute) for helpful discussions and critical II to promoter sequences (refs. 16 and 17 and references reading of the manuscript, and we are grateful to Sara Windahl and therein), and subsequently, TFIIF increases elongation effi- Tova Almlo¨f (Department of Biosciences, Karolinska Institute) for ciency by preventing pausing and so reducing arrest of the reagents. This work was supported by a grant from the Swedish polymerase molecule (refs. 16, 17, and 39 and references Medical Research Council (13X-2819). therein). While the present results are compatible with a role for the AR in preinitiation complex assembly, it remains 1. Zhou, Z. X., Wong, C. I., Sar, M. & Wilson, E. M. (1994) Recent Prog. Horm. Res. 49, 249–274. possible that the receptor also regulates at the 2. Quigley, C. A., De Bellis, A., Marschke, K. B., el-Awady, M. K., level of elongation. Recently, a number of groups have dem- Wilson, E. M. & French, F. S. (1995) Endocr. Rev. 16, 271–321. onstrated a role for activator proteins in elongation and 3. Brinkmann, A. O., Jenster, G., Ris-Stalpers, C., van der Korput, correlated this to interactions with another of the basal H., Bruggenwirth, H., Boehmer, A. & Trapman, J. (1996) transcription factors, TFIIH (40–43). Interestingly, prelimi- 61, 172–175. nary results from kinetic studies of the AR-dependent squelch- 4. Trifiro, M. A., Kazemi-Esfarjani, P. & Pinsky, L. (1994) Trends ing suggest that the receptor can act after preinitiation com- Endocrinol. Metab. 5, 416–421. plex assembly (unpublished observations), and further inves- 5. Brown, T. R. (1996) Prostate Suppl. 6, 9–12. tigation will be required to resolve these possible roles of the 6. Klocker, H., Culig, Z., Hobisch, A., Cato, A. C. & Bartsch, G. Prostate 25, AR in transcription regulation. (1994) 266–273. 7. Bentel, J. M. & Tilley, W. D. (1996) J. Endocrinol. 151, 1–11. Although many studies have focused on the general tran- 8. Lobaccaro, J.-M., Lumbroso, S., Belon, C., Galtier-Dereure, F., scription factors TBP and TFIIB as possible targets for acti- Bringer, J., Lesimple, T., Heron, J.-F., Pujol, H. & Sultan, C. vator proteins, including nuclear receptors (refs. 16, 22, and 44 (1993) Nat. Genet. 5, 109–110. and references therein), a number of activators, including SRF 9. Brinkmann, A. O., Faber, P. W., van Rooij, H. C., Kuiper, G. G., (27), the protooncogenes c-myc (24), c-fos, and c-jun (45), and Ris, C., Klaassen, P., van der Korput, J. A., Voorhorst, M. M., van the arylhydrocarbon receptor and receptor transporter pro- Laar, J. H., Mulder, E. & Trapman, J. (1989) J. Steroid Biochem. teins (23), have been shown to interact with TFIIF subunits. 34, 307–310. While the functional significance of such interactions has not 10. Jenster, G., van der Korput, H. A. G. M., van Vroonhoven, C., always been possible to demonstrate, the integrity of the van der Kwast, T. H., Trapman, J. & Brinkmann, A. O. (1991) Mol. Endocrinol. 5, 1396–1404. RAP74 subunit was found to be critical for SRF-dependent 11. Simental, J. A., Sar, M., Lane, M. V., French, F. S. & Wilson, transactivation of a reporter gene (27). E. M. (1991) J. Biol. Chem. 266, 510–518. In recent years there have been a plethora of protein– 12. Palvimo, J. J., Kallio, P. J., Ikonen, T., Mehto, M. & Ja¨nne,O. A. protein interactions described involving members of the nu- (1993) Mol. Endocrinol. 7, 1399–407. clear receptor superfamily and basal transcription factors, 13. Jenster, G., de Ruiter, P. E., van der Korput, H. A., Kuiper, co-activator proteins, and proteins of unknown function (for G. G., Trapman, J. & Brinkmann, A. O. (1994) Biochemistry 33, review see ref. 22 and references therein). In the majority of 14064–14072. cases these interactions have involved the generally weaker 14. Jenster, G., van der Korput, H. A., Trapman, J. & Brinkmann, transactivation functions (AF-2, AF-␶c) localized in the C- A. O. (1995) J. Biol. Chem. 270, 7341–7346. 15. Chamberlain, N. L., Whitacre, D. C. & Miesfield, R. L. (1996) terminal ligand-binding domain (22). While many of the J. Biol. Chem. 271, 26772–26778. described interactions occur with several different receptors, 16. Conaway, R. C. & Conaway, J. W. (1993) Annu. Rev. Biochem. 62, Yeh et al. (46), using a two-hybrid screen, identified a protein 161–190. (ARA70) that showed specificity for the androgen receptor 17. Orphanides, G., Lagrange, T. & Reinberg, D. (1996) Genes Dev. steroid-binding domain. In the present study we demonstrate 10, 2657–2683. that a region of the N terminus, encompassing the AF-1 18. Roeder, R. G. (1996) Trends Biochem. Sci. 21, 327–335. function of the human AR, acts to recruit the general tran- 19. Stargell, L. A. & Struhl, K. (1996) Trends Genet. 12, 311–315. scription machinery to a target promoter. In addition, this 20. Kaiser, K. & Meisterernst, M. (1996) Trends Biochem. Sci. 21, region of the receptor selectively interacts with the subunits of 342–345. 21. Verrijzer, C. P. & Tjian, R. (1996) Trends Biochem. Sci. 21, the general transcription factor TFIIF. This report of an 338–342. interaction involving the major transactivation function of the 22. McEwan, I. J., Wright, A. P. H. & Gustafsson, J.-Å. (1997) human AR indicates possible mechanisms whereby the DNA- BioEssays 19, 153–160. bound receptor activates gene transcription. 23. Rowlands, J. C., McEwan, I. J. & Gustafsson, J.-Å. (1996) Mol. Pharmacol. 50, 538–548. Note Added in Proof. Recently, Snoek et al. (47) have used nuclear 24. McEwan, I. J., Dahlman-Wright, K., Ford, J. & Wright, A. P. H. extracts derived from the prostate carcinoma cell line PC3, together (1996) Biochemistry 35, 9584–9593. with glutathione S-transferase (GST) fusion proteins to reconstitute 25. McEwan, I. J., Wright, A. P., Dahlman-Wright, K., Carlstedt- rat AR function in vitro. In this system a recombinant protein encoding Duke, J. & Gustafsson, J.-Å. (1993) Mol. Cell. Biol. 13, 399–407. the receptor DNA-binding domain (DBD) and steroid-binding do- 26. McEwan, I. J., Almlo¨f, T., Wikstro¨m, A.-C., Dahlman-Wright, K., main (SBD) activated a reporter gene more robustly than a fusion Wright, A. P. H. & Gustafsson, J.-Å. (1994) J. Biol. Chem. 269, protein containing a region of the N terminus and DBD. Although this 25629–25636. latter construct overlaps with the region used in the present study, a 27. Joliot, V., Demma, M. & Prywes, R. (1995) Nature (London) 373, direct comparison is not possible, as we have used a heterologous 632–635. DBD, to avoid the possible influence of DNA-binding efficiency, by 28. De Vos, P., Schmitt, J., Verhoeven, G. & Stunnenberg, H. G. the receptor DBD, on transactivation. (1994) Nucleic Acids Res. 22, 1161–1166. 29. Sawadogo, M. & Roeder, R. G. (1985) Cell 43, 165–175. We gratefully acknowledge the gift of plasmids from the following 30. Lue, N. F., Flanagan, P. M., Sugimoto, K. & Kornberg, R. D. people: Drs. A. Berk (University of California at Los Angeles), A. O. (1989) Science 246, 661–664. Downloaded by guest on October 2, 2021 8490 Biochemistry: McEwan and Gustafsson Proc. Natl. Acad. Sci. USA 94 (1997)

31. Ptashne, M. & Gann, A. A. F. (1990) Nature (London) 346, 1419–1428. 329–331. 40. Yankulov, K., Blau, J., Purton, T., Roberts, S. & Bentley, D. L. 32. Berger, S. L., Cress, W. D., Cress, A., Triezenberg, S. J. & (1994) Cell 77, 749–759. Guarente, L. (1990) Cell 61, 1199–1208. 41. Blair, W. S., Fridell, R. A. & Cullen, B. R. (1996) EMBO J. 15, 33. Kelleher, R. D., Flanagan, P. M. & Kornberg, R. D. (1990) Cell 1658–1665. 61, 1209–1215. 42. Blau, J., Xiao, H., McCracken, S., O’Hare, P., Greenblatt, J. & 34. Greenblatt, J. (1991) Trends Biochem. Sci. 16, 408–411. Bentley, D. (1996) Mol. Cell. Biol. 16, 2044–2055. 35. Kim, Y. J., Bjorklund, S., Li, Y., Sayre, M. H. & Kornberg, R. D. 43. Yankulov, K. Y., Pandes, M., McCracken, S., Bouchard, D. & (1994) Cell 77, 599–608. Bentley, D. L. (1996) Mol. Cell. Biol. 16, 3291–3299. 36. Koleske, A. J. & Young, R. A. (1994) Nature (London) 368, 44. Zawel, L. & Reinberg, D. (1995) Annu. Rev. Biochem. 64, 466–469. 533–561. 37. Ossipow, V., Tassan, J. P., Nigg, E. A. & Schibler, U. (1995) Cell 45. Martin, M. L., Lieberman, P. M. & Curran, T. (1996) Mol. Cell. 83, 137–146. Biol. 16, 2110–2118. 38. Maldonado, E., Shiekhattar, R., Sheldon, M., Cho, H., Drapkin, 46. Yeh, S. & Chang, C. (1996) Proc. Natl. Acad. Sci. USA 93, R., Rickert, P., Lees, E., Anderson, C. W., Linn, S. & Reinberg, 5517–5521. D. (1996) Nature (London) 381, 86–89. 47. Snoek, R., Rennie, P. S., Kaspar, S., Matusik, R. J. & Bruchovsky, 39. Aso, T., Conaway, J. W. & Conaway, R. C. (1995) FASEB J. 9, N. (1996) J. Steroid Biochem. Mol. Biol. 59, 243–250. Downloaded by guest on October 2, 2021